Have a personal or library account? Click to login

Is Google Gemini better than ChatGPT at evaluating research quality?

Open Access
|May 2025

Figures & Tables

Figure 1.

Spearman correlations between Gemini 1.5 Flash scores and the author’s scores for 51 library and information science articles, against the number of repetitions averaged. Each line represents a different amount of input. Error bars are 95% confidence intervals for averaging within the data collected.
Spearman correlations between Gemini 1.5 Flash scores and the author’s scores for 51 library and information science articles, against the number of repetitions averaged. Each line represents a different amount of input. Error bars are 95% confidence intervals for averaging within the data collected.

Figure 2.

Spearman correlations between Gemini 1.5 Flash scores and departmental average REF2021 scores. Also shown are equivalent correlations from ChatGPT 4o-mini and, as a benchmark, the correlation between article scores and departmental average REF2021 scores. Error bars are 95% confidence intervals for the assumed infinite population of similar articles.
Spearman correlations between Gemini 1.5 Flash scores and departmental average REF2021 scores. Also shown are equivalent correlations from ChatGPT 4o-mini and, as a benchmark, the correlation between article scores and departmental average REF2021 scores. Error bars are 95% confidence intervals for the assumed infinite population of similar articles.
DOI: https://doi.org/10.2478/jdis-2025-0014 | Journal eISSN: 2543-683X | Journal ISSN: 2096-157X
Language: English
Page range: 1 - 5
Submitted on: Dec 9, 2024
Accepted on: Dec 25, 2024
Published on: May 6, 2025
Published by: Chinese Academy of Sciences, National Science Library
In partnership with: Paradigm Publishing Services
Publication frequency: 4 times per year

© 2025 Mike Thelwall, published by Chinese Academy of Sciences, National Science Library
This work is licensed under the Creative Commons Attribution 4.0 License.